BUSINESS INTELLIGENCE FOR DECISION SUPPORT SYSTEMS P LENIN TECHNICAL DIRECTOR NATIONAL INFORMATICS CENTRE MINISTRY OF COMMUNICATION & IT NEW DELHI -110 003. 1 Why Now, Not Then? DATA AVAILABILITY –MASSIVE AMOUNT /ELECTRONIC FORM –WIDESPREAD USE OF APPLN SYSTEMS –POWERFUL H/W, PARALLEL SYSTEMS –NEW S/W TECHNOLOGY 2 DEMAND FOR DATA Growing gap Source: Gartner Group 3 ER MODEL P&V DB Admin -I District Admin -II States Appln Div. Purchas e Projects Web Services DG Office Library NICSI ABC Div. A/C Div User Min.s XYZ Div. Stores Div Contact LMN Division 4 WHY BUSINESS INTELLIGENCE SOLUTIONS ? BI SYSTEMS NOT A REPLACEMENT FOR RDBMS/OLTP TO COMPLEMENT RDBMS/OLTP SYSTEMS FOR PLANNERS AND DECISION MAKERS * USE WHEREVER AND WHENEVER /WEB BASED * MINIMUM DEPENDENCY ON IT PROFESSIONAL * EXTRACT IMPLICIT KNOWLEDGE/PATTERN 5 WHY BI/ DATA WAREHOUSE? -NEED OF INFORMATICS REQUIREMENT IS DYNAMIC -ALSO NEED OF EVERY DEPTT VARIES [SECRETARY/COMMISSIONER/JT.COMM/DY.COMM.] -INTEGRATED/SHARED INFORMATION •DESIGN NOT SUITABLE FOR BUSINESS ANALYSIS • LIMITED ANALYSIS • DATA CONTENT CHANGES • VIEWS BASED ON OBJECTIVES • REASONING [EX. PDS PUNJAB/ORISSA] 6 WHY BI TECHNOLOGY? - TO HANDLE DATA FROM MUTI-SOURCE - TO HANDLE LARGE VOLUME OF DATA - MULTIDIMENSIONAL –DRILL THRU-ANALYSIS - PRESERVE AND USE THE HISTORICAL DATA - TO EXTRACT HIDDEN KNOWLEDGE/PATTERN - TO DO AD HOC QUERY : : 7 WHAT BENEFITS YOU GET OUT OF DW? * EASIER FOR END USERS TO NAVIGATE, UNDERSTAND AND QUERY • ENABLE QUERIES THAT CUT ACROSS DIFFERENT SEGMENTS • COMPLEX QUERY IN NORMALIZED DB CAN BE BUILD EASILY • EFFICIENT WAY TO MANAGE AND REPORT • PROVIDES THE CAPABILITY TO ANALYZE LARGE AMOUNT OF HISTORICAL DATA 8 Data Warehouse W H INMON, A DW IS A SUBJECT ORIENTED (Vs APPLN. ORIENTED), INTEGRATED, TIME VARIANT (HISTORICAL vs. CURRENT), NONVOLATILE (STABLE vs. CONTINUOUS CHANGE) COLLECTION OF DATA IN SUPPORT OF MANAGEMENT‘S DECISION MAKING PROCESS. Maintained separately from the organization’s operational database 9 A SAMPLE DATA CUBE /DIMENSION MODEL 2Qtr 3Qtr 4Qtr sum Total of all product sales by country and quarter Total annual sales by country and product U.S.A CHINA INDIA Country TV PC DVD sum 1Qtr Date Total of all item sales in all countries by product sum Total annual sales of all items by Country Total annual sales. Total of all product sales in all countries by quarter 10 DATE ORDER NO 08/08/04 1 2 3 4 5 09/08/04 CARS SOLD SANTRO 5 3 2 2 3 MARUTI 2 0 6 2 3 1 3 7 2 2 4 1 3 0 11 CARS SOLD Date Santro Maruti 08/08/04 15 13 09/08/04 9 8 NO OF RECORDS REDUCED BY AGGREGATION 12 Transformation of Data Information Exploration / analysis SQL reporting Warehouse Cleansing / normalization Data Transaction processing 13 Basic elements of Data warehouse OPERATIONAL SOURCE SYSTEM DATA STAGING AREA DATA PRESENTATION AREA Services: Extract Extract Clean, combine, and standardize Conform Dimensions No user query services DATA ACCESS TOOLS Data Mart #1 Load Dimensional Atomic and summary data Based on a single business process Ad hoc query tools Access Report Writers Analytical Applications Data Store: DW Bus: Flat files and relational tables Conformed facts and dimensions Modeling: Processing: Extract Sorting and sequential processing Forecasting Data Mart #2 Scoring Access Load Similar design Data Mining 14 METADATA METADATA 15 THE OLAP SOLUTION A PROCESS OF FAST ANALYSIS OF SHARED MULTIDIMENSIONAL INFORMATION --ENTERPRISE-WIDE DATA ANALYSIS 16 OLAP - FAST ? - WHAT ANALYSIS? - SHARED - MULTIDIMENSIONAL ON ENTERPRISE-WIDE DATA VIEW FROM DIFFERENT ANGLE 17 ** What all with OLAP? # Drill-down Process (Browsing) Time LOCATION SOURCE Year REGION SALES TAX Quarter STATE CEN.EX Month District VAT Week City XYZ 18 BI SYSTEMS DEVELOPMENT IN NIC • SOCIO ECONOMIC DATA OF UP • BHOOMI LAND RECORDS • IRRIGATION SURVEY DATA • SPECIAL FRAUD INVESTIGATION • FERTILIZER PROD/MOVT/CONSUMPTION • EXPORT/IMPORT OF MAJOR COMM+INDEXES FOR RIS/COMMERCE MIN. * CUSTOMS IMPORT --- ALL AT NIC HQR. 19 BI SYSTEMS DEVELOPMENT IN NIC STATE UNITS: • KERALA Treasury Expenditure Sub-Reg Office Stamp Duty Collection Civil Supplies Office Expense Disease survelliance • TAMIL NADU Child Labour Education • MAHARASHTRA/PUNE • ANDHRA PRADESH • GUJARAT 20 CUSTOMS IMPORT DATA LIMITED INPUT CUSTOMS IMPORT DATA ITEM LEVEL DATA [Bill of Entry, Assessment value, Duty, Duty Foregone, CTH, CETH, BCD NOTFN, ….] 20 SITES ONLY TIME RANGE: APRIL 2003 TO DEC. 2004 DATA SIZE = 1.45 Cr RECORDS = 5GB 21 NIC DW ARCHITECTURE RETRIEVE STORAGE AREA NETWORK (SAN) Extract DATA SERVER (ORACLE 9i/10G) Store ETL (DECISION STREAM) Load DATA SOURCE CLIENT Request COGNOS SERVER POWERPLAY REPORTNET To access & change the layout of Powerplay and ReportNet reports… CLEMENTINE Response WINDOWS 2000 SERVER (OS) 22 BI FOR BHOOMI LAND RECORDS - TO DEMONSTRATE THE POWER OF BI TECH - JOINT EFFORT BY A&M DIV AND NIC-KSU - NIC KSU …DOMAIN EXPERT - A&M DIV AS BI TECH EXPERT - FOR KANNADA LANGUAGE TOOK ASSISTANCE FROM KEYSOFT 23 BI SYSTEM PLAN FOR KSU STATE • USE THE HW/SW FACILITIES AT THE HQRS. • KEEP THE DATA AT KSU SERVER AND USE HQR FACILITIES FOR SW • HAVE BOTH HW AND SW IN THE KSU SERVER 24